知识卡片
时间本身可以改变域
季节、年份和人的行为变化,都可能让输入或目标关系偏离过去数据。时间切分与按部署场景分层的评估可暴露部分变化;随机切分常让训练和测试共享历史条件。监测应围绕实际使用时段和环境,而非默认未来保持历史分布。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
原书出处 · 1 处
- · 查看原书依据
Aside from colorizing photos on Reddit, how often do situations like this come up in the real world? All the time. Suppose you train a weather prediction system on data from the winter months, then test it in the summer. Or train it in 2020 and test it in 2030. What would happen if there were global warming? Suppose you trained a predictive model at
参考来源
- 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第35章《数据偏差与分布漂移》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c035b.xhtml)
- 结论依据:摘录中关于时间本身可以改变域的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:Aside from colorizing photos on Reddit, how often do situations like this come up in the real world? All the time. Suppose you train a weather prediction system on data from the winter months, then test it in the summer. Or train it in 2020 and test it in 2030. What would happen if there were global warming? Suppose you trained a predictive model at